JP5208972B2 - Signal quality determination in cable networks - Google Patents

Description

  The present invention relates to the determination of signal quality in a cable network. In particular, the present invention relates to a method and apparatus for determining signal quality in a cable network using a network model.

  It is well known to determine signal quality parameters of electrical and optical networks, including cable networks such as CATV (Cable television) networks. Parameters or criteria that are indicators of signal quality at the receiving end (usually the subscriber end) are, for example, the signal to noise ratio (SNR) and the bit error rate (BER) in the digital network. By determining the noise level (and / or other characteristics) of the output signal, an indication of signal quality at the receiving end of the network can be obtained.

  The cable network includes a plurality of cables, a plurality of amplifiers, and other components. Although the impact on cable or wiring signal quality is relatively small, they attenuate high frequency components. Amplifiers usually generate noise. The noise of many amplifiers connected in series, as used in normal cable networks, accumulates and affects the actual signal.

In addition, the amplifier generates signal distortion due to non-linearity. Ideally, the amplifier outputs an input signal s (t) _in multiplied by a gain factor.
s (t) _out = A · s (t) _in
Here, the gain A is a constant. In practice, however, the amplifier is not perfectly linear, and the output signal usually contains higher order series terms of the input signal, including second and third order terms.
s (t) _out = A · s (t) _in + B · s (t) _in ² + C · s (t) _in ³ + ...
As a result of this non-linearity, so-called intermodulation of the input signal frequency results in an output signal containing frequency components not present in the input signal. These undesirable intermodulations degrade signal quality and must be taken into account when determining network signal quality.

  However, conventional methods generally cannot take these intermodulations into account. Even when intermodulation is taken into account, each intermodulation is naturally united and individual contributions to the overall signal quality level cannot be accurately determined.

  The paper “Frequency Response of Non-linear Networks using Curve Tracing Algorithm” (A. Yoshida, Y. Yamagami, & Y. Nishio, IEEE, May 2002) A method for calculating a characteristic curve of a circuit is disclosed. Non-linear effects are taken into account, but only the fundamental wave component is taken into account. Therefore, the effect of intermodulation is considered at only one frequency, and this known method is unsuitable for determining the intermodulation effect of multiple frequency bands. Furthermore, the prior art paper does not mention cable networks.

  It is an object of the present invention to provide a method and apparatus for overcoming these and other problems of the prior art and determining signal quality in cable networks that yields more accurate results.

The present invention provides a method for determining signal quality in a network, which method comprises the following steps.
Providing a network model including an interconnect and a model of at least one element; providing an input signal; determining an output signal using the input signal and the network model; using the input signal and the output signal. Determining the signal quality criteria The method has the following characteristics.
-The network model is a model of a cable network such as a cable TV network.-The input signal includes the component signal components expressed in multiple frequency domains.-The output signal is the output signal of the cable network in the frequency domain. The step of determining the output signal includes the following: ○ Simulate the behavior of the cable network according to the input signal using the network model ○ By convolving the spectrum of the component signal components・ Determine frequency domain intermodulation ○ Create signal quality standards using this frequency domain intermodulation and output signal

  By using an input signal having multiple component signal components, the contribution of these components to intermodulation and thereby the contribution to signal quality can be accurately determined. By further simulating the behavior of the cable network in response to an input signal having a plurality of constituent signal components, very practical results can be obtained.

  By determining the frequency domain intermodulation by performing convolution of multiple spectra of the constituent signal components, both single frequency signal components and frequency components with non-zero bandwidth can be processed. The ability to determine intermodulation of signals having multiple signal components or multiple non-zero bandwidths is clearly advantageous over the prior art.

  Furthermore, the use of frequency domain intermodulation ensures that, in addition to the output signal, the signal quality criterion reliably considers both the desired output signal and the intermodulation to create a signal quality criterion. As a result, a very reliable signal quality criterion is obtained, which can be used in a wide range of input signals and a wide range of cable networks.

  The intermodulation resulting from the constituent signal components is preferably determined separately by performing individual convolutions, so that the effect of the intermodulation on the signal quality of the network can be accurately determined.

  As described above, frequency domain intermodulation can be determined by convolution of the spectrum of the constituent signals. However, actual convolution requires a large processing capacity. Thus, each convolution of the spectrum of the constituent signals is accomplished by performing an inverse Fourier transform, a time domain multiplication, and a Fourier transform. In other words, frequency domain convolutions are performed effectively, but are even more effective than convolutions because these convolutions are performed substantially in time domain multiplication. As is well known, forward and backward (ie, inverse) Fourier transforms can be performed very effectively using FFT (Fast Fourier Transform).

  The network model is preferably a frequency domain model, and has a component that directly affects the spectrum of a plurality of signals. This has the advantage that signal characteristics in the frequency domain can be used. A signal used in a cable network such as a CATV network is usually specified in the frequency domain by indicating the center frequency and bandwidth of the signal, for example, and such specification determines the input signal of the model. Used directly.

  In one preferred embodiment, the element model comprises a gain unit, a weighting unit, and at least one intermodulation unit, each of which includes gain contribution, frequency dependence, and intermodulation of the element. Used to determine the contribution of. Such an element model makes it possible to accurately model the characteristics of the element.

  More preferably, the element model comprises at least two intermodulation units and determines second-order and third-order intermodulation, respectively. The element model can be used to consider only the second order intermodulation or only the third order intermodulation, but including both the second and third order intermodulation can greatly improve the modeling. Fourth order or higher order intermodulation can also be modeled using higher order intermodulation units, however, the resulting computational complexity is much higher than the resulting computational complexity. There is no effect.

  Preferably, the at least one intermodulation unit has a primary weighting unit that performs weighting of an input signal before determining intermodulation in the preceding stage, and 2 that performs intermodulation weighting in the subsequent stage. There are the following weighting units: Although it is possible to use only one weighting unit for each branch, a more accurate model is obtained by using weighting units before and after the intermodulation unit.

  The at least one intermodulation unit preferably comprises a plurality of intermodulation subunits that determine the intermodulation of the constituent signal components.

  The present invention also provides a computer program product for performing the above method. Computer program products comprise computer-executable instructions stored in a data medium such as a CD or DVD. A set of computer-executable instructions can cause a programmable computer to execute the above method, and can be obtained by downloading from a remote server via the Internet, for example.

The invention further provides an apparatus for determining signal quality in a network, the apparatus comprising the following units:
A memory unit that stores a network model including an interconnection and at least one element model an input unit that provides an input signal a processing unit that determines an output signal using the input signal and the network model A signal quality unit that determines the signal quality criteria to use. The apparatus has the following characteristics:
The network model stored in the memory unit is a model of a cable network such as a cable television network. The input unit is configured to receive an input signal including a component signal component represented in a multi-frequency domain. The processing unit is configured to determine an output signal that includes a representation of the output signal of the cable network in the frequency domain. The processing unit further determines the output signal by: ○ Simulate the behavior of the cable network according to the input signal using the network model ○ Decide the frequency domain intermodulation by convolving the spectrum of the component signal components ○ This frequency signal quality by using the intermodulation output signal region To generate a reference

  The invention will be further described below with reference to the embodiments shown in the accompanying drawings.

FIG. 1 schematically shows an example of a network model used in the present invention. FIG. 2 schematically shows an element model according to the invention. FIG. 3 schematically shows a first intermodulation unit according to the invention. FIG. 4 schematically shows a second intermodulation unit according to the invention. FIG. 5 schematically shows a method and settings for determining signal quality in a cable network according to the invention. FIG. 6 schematically illustrates a method and configuration for updating a network model according to the present invention.

  The network model 1 shown in the embodiment of FIG. 1 does not limit the present invention to this. The network model 1 includes an interconnection 2, an element model 3, an input terminal 4, and an output terminal 5. Network model 1 is composed of three interconnected amplifiers and represents an actual cable network (not shown) having one input terminal and two output terminals.

  Such a network model is well known so that the cable manager can determine the signal quality at the output terminal 5. This model provides an indication of the noise contribution of the amplifier in the presence of an input signal to the network. The signal quality determined at the output terminal 5 indicates the quality of service received by the subscriber.

Conventional methods are often based on frequency-independent linear amplification models, which do not take into account all the effects of frequency-dependent characteristics and amplifier nonlinear characteristics. This is particularly problematic when wideband and / or multiple input signals are used. This is the case typical of modern cable networks. Any nonlinearity in the amplifier will produce an intermodulation component. That is, a new frequency component is generated from nonlinear amplification of the input signal. For example, the input frequencies f ₁ and f ₂ give rise to additional frequencies f ₁ + f ₂ and f ₁ −f ₂ when using standard amplifiers. These additional signal components are undesirable and contribute to the overall noise level of the output signal.

  However, conventional network models generally assume that the amplifier is completely or approximately linear and therefore do not take into account any additional noise due to intermodulation. From this, the noise prediction may result in significantly lower than the actual noise level. As a result, the quality of service received by the subscriber is lower than expected. The present invention solves this problem using an improved element model.

  An element model according to the invention is schematically shown in FIG. Element model 3, which is merely an example, includes gain (G) or linear amplification unit 31, intermodulation (IM) or nonlinear amplification units 32 and 33, first-order weighting (PW) units 34, 35 and 36, 2 The following weighting units 37 and 38 are provided. The element model 3 receives the input signal IS and outputs an output signal OS. At least one of the IS and OS signals may be identical to the respective counterpart IS or OS of FIG. 1, but this is not a requirement.

  The gain (G) unit 31 models a linear gain of a network element, and typically models an amplifier. This gain is frequency independent. A first primary weighting (PW1) unit 34 performs weighting in the frequency domain of the input signal IS so that one frequency is attenuated from the other frequency. This feature makes it possible to model the frequency-dependent transmission characteristics of network elements, and thus to model an actual network. As is well known, in cable networks, signal attenuation usually increases with frequency.

  A second order intermodulation (IM2) unit 32 determines the intermodulation resulting from the second order terms in the amplification characteristics (ie, generally transmission characteristics) of the network element. The secondary intermodulation unit 32 has a second primary weighting (PW2) unit 35 at the front stage and a second secondary weighting (SW2) unit 37 at the rear stage. These two weighting units are respectively input. Frequency-dependent weighting of signal IS and second order intermodulation are performed. The weighting unit 37 outputs a second-order intermodulation signal IMS2.

  It is possible to perform frequency-dependent weighting using only one weighting unit before or after the intermodulation unit 32. However, according to a further aspect of the invention, it is preferred to have both primary and secondary weighting units. In this way, better weighting and more accurate network element modeling is achieved.

  In the embodiment of FIG. 2, only one weighting unit 34 is arranged in series with the gain unit 31, but this has no effect even if two weighting units are arranged in the linear branch of the element model. Because there is no. In this embodiment, therefore, there is only one (first) primary weighting (PW1) unit 34 and the secondary weighting unit is omitted. It is understood that there may be a secondary weighting unit instead of the primary weighting unit 34.

  A third order intermodulation (IM3) unit 33 determines the intermodulation resulting from the third order term in the amplification characteristics (ie, generally transmission characteristics) of the network element. This third-order intermodulation unit 33 has a third primary weighting (PW3) unit 36 at the front stage and a third secondary weighting (SW3) unit 38 at the rear stage, and these two weighting units are respectively input. Frequency-dependent weighting of signal IS and third-order intermodulation are performed. The weighting unit 38 outputs a third-order intermodulation signal IMS3.

  As before, it is possible to perform frequency-dependent weighting using only one weighting unit before or after the intermodulation unit 33, but according to a further aspect of the invention, the third order of the model Preferably, both weighting units are used in the intermodulation branch.

The intermodulation units 32 and 33 will now be described in more detail with reference to FIGS. In accordance with an important aspect of the present invention, the input signal used (IS in FIG. 1) includes multiple components, each component representing a signal type. For example, the input signal consists of two or more of the following components.
・ PAL (Phase Alternating Line): TV signal ・ FM (Frequency Modulation): Radio signal ・ QAM (Quadrature Amplitude Modulation): Data communication ・ SPAL: (synchronized PAL): TV signal ・ OFDM (Orthogonal Frequency Division Multiplexing): Data communication Carrier: network measurement and control signals. These input signals are provided in a frequency domain (ie spectrum) representation. In the examples of FIGS. 3 and 4, only two input signal components P and Q are shown, but in practice more than two input components may be used.

The (second order) intermodulation unit 32 of FIG. 3 includes intermodulation subunits 321, 322, 323 to determine intermodulation of a plurality of component signal components. The first subunit 321 receives only the signal component P and generates its own intermodulation with the component P, indicated by the symbol x _P ² , to generate the intermodulation component PP. Similarly, the third sub-unit 323 receives only the signal component Q, is indicated by symbol x _Q ^2, to produce the intermodulation component Q with itself, it generates intermodulation component QQ.

The second subunit, however, receives both signal component P and signal component Q and performs a “true” intermodulation of component P and component Q, indicated by the symbol x _{P ·} x _Q. To generate an intermodulation component PQ. Thus, the intermodulation of the plurality of component signal components is determined separately by the plurality of subunits. By determining the intermodulation components separately, a very accurate representation of the intermodulation is obtained, and thus a very accurate signal quality prediction.

  Since the input signal (IS in FIG. 1) is provided as a spectrum (frequency domain representation), the constituent signal components P, Q and the intermodulation signals PP, PQ, QQ are signal representations in the frequency domain.

The (third order) intermodulation unit 33 shown in FIG. 4 comprises intermodulation subunits 331, 332, 333, 334 so as to determine the intermodulation of the constituent signal components P and Q. The first subunit 331 receives only the signal component P, generates the component P and its own (third order) intermodulation, indicated by the symbol x _P ³ , and generates the intermodulation component PPP To do. Similarly, the fourth subunit 334 receives only the signal component Q and generates the intermodulation component QQQ by generating the component Q and its own intermodulation, indicated by the symbol x _Q ³ .

  The second subunit 332, however, receives both the signal component P and the signal component Q and generates an intermodulation component PPQ. Similarly, the third subunit 333 generates an intermodulation component PQQ.

It can be seen that the intermodulation unit 32 determines separate intermodulation components PPP, PPQ, PQQ, QQQ from the input signal components P and Q. As described above, the signal components P and Q are frequency domain signals, that is, more specifically, representations of the frequency domain of the time signal. The products x _P ³ , x _P ² · x _Q, etc. are time domain products, which are calculated in the frequency domain using a convolution process that requires a large amount of computation. For this reason, the units 32 and 33 are preferably used to convert (inverse) the frequency domain signal components P, Q into the time domain, and the time domain products x _P ³ , x _P ² .x _Q etc. The frequency domain intermodulation components PP,. . . , QQ or PPP,. . . , Comprising a fast Fourier transform (FFT) unit for obtaining QQQ.

  The network model (1 in FIG. 1), the element model (3 in FIG. 2), the intermodulation units 32 and 33, and the subunits thereof are implemented by hardware, software, or a combination thereof. Good. The software is preferably adapted to run on a conventional computer system.

  The determination of signal quality according to the present invention is illustrated in FIG. The network model 1 'represents a cable network with two amplifiers. The element model 3 corresponding to this has three outputs, as shown in FIG. 2, and generates an output signal OS, a second-order intermodulation signal IMS2, and a third-order intermodulation signal IMS3, respectively. The output signal of the first element model 3 is supplied to the second element model for amplification, while this intermodulation signal is supplied to gain (G) adjustment units 302 and 303. The gain-adjusted first element model intermodulation signals are added to the second element model intermodulation signals in summing units 304 and 305, respectively, to produce combined intermodulation signals IMS2 and IMS3.

  Gain adjustment units 302 and 303 are shown as separate units for adjusting the second order intermodulation gain (G2) and the third order intermodulation gain, respectively. In other embodiments, a single combined gain adjustment unit may be used. The gains of the gain adjustment units 302 and 303 correspond to the gains of all other elements (amplifiers) in the network model. In this embodiment, the gain adjustment units 302 and 303 have a gain equal to the gain of the second amplifier model 3. Units 302 and 303 preferably perform frequency adjustment or frequency weighting in addition to gain adjustment. This weight is equal to the weight of all other element models. Therefore, in this embodiment, the intermodulation is weighted as if it had passed through the second element model 3 (the gain is also adjusted).

  The output signal OS of the second element model, the second order intermodulation signal IMS2, and the third order intermodulation signal IMS3 are separately supplied to a signal quality (SQ) unit 309, which in this embodiment is the signal quality unit. Generate a signal-to-noise ratio (SNR) and a bit error rate (BER) of the signal.

  The intermodulation signals IMS2 and IMS3 are respectively composed of intermodulation signals constituting them, for example, the configuration signals PPP, PPQ, etc. of FIG. In the signal quality unit 309, the constituent signals are added separately. That is, the PPP contributions from both amplifier models 3 in FIG. 5 are added to form a combined PPP contribution, the PPQ contributions are added to form a combined PPQ contribution, etc. is there. Next, using the combined contribution, the output signal OS, and the specifications of the input signals (eg, QAM, PAL, and FM signals) used to generate the input signal IS, the SNR and / or BER is Calculated. These input signal specifications (ISS) are included in the stored list of specifications 9 and may include (carrier) frequency, signal level, bandwidth, and other parameters. Deriving the input signal IS (in the frequency domain) from the input signal specification ISS in list 9 will be described later with reference to FIG.

  In addition to the effect of intermodulation on the signal quality level, a noise model may additionally be used. A conventional noise model that assumes thermal noise as the model input may be used. The gain and weighting characteristics of the element model selectively include any form of noise that represents the noise caused by that element to determine the output noise level contained in the output signal OS provided to the signal quality unit 309. Used.

  The processing shown in FIG. 5 is preferably processed by software, but may be implemented by hardware.

  The multiple element models (3 in FIGS. 1, 2 and 5) include parameters, such as gain parameters and weighting parameters. These parameters can be determined using the configuration of FIG. 6, but this may be implemented in software and / or hardware.

  In the element model unit 3, this model shown in the present embodiment is an amplifier and receives model parameters (pars). These parameters are generated by a parameter adjustment (PA) unit 7, which will be described later. The element model unit 3 receives an input signal IS (in the frequency domain) from an input signal generation (ISG) unit 8, and the input signal generation unit 8 receives an input signal specification (ISS) from a stored input signal specification list 9. . As mentioned above, the input signal specification may include (carrier) frequency, bandwidth, power level, and / or other parameters. The input signal used in model 3 may be a set of digital data representing a physical input signal, or it may be an actual digital input signal.

  The input signal generation (ISG) unit 8 generates the frequency domain input signal IS using the input signal specifications. This input signal specification is (center) frequency, bandwidth, output level, (spectral) envelope, and / or other parameters. Signal generators that can generate an input signal based on these and similar parameters are well known.

  The input signal specification (ISS) is also provided to the second input signal generation (ISG) unit 8 '. The second input signal generation (ISG) unit 8 'generates a physical (frequency domain) input signal IS' which is fed to the actual element (in this embodiment an amplifier) 3 '. It is done.

  The model unit 3 outputs a composite output signal including a basic output signal (OS) and intermodulation signals IMS2 and IMS2. Similarly, the element unit 3 outputs a composite signal including a basic output signal (OS ') and intermodulation signals IMS2' and IMS3 '. These signals are received and compared by the comparison unit 6. Any difference between the calculation signal generated by the model 3 and the measurement signal generated by the actual element 3 ′ becomes the difference signal DS and is given to the parameter adjustment unit 7.

The parameter adjustment unit 7 determines the model parameters of the element model, in particular the weighting parameters of the weighting units (34-36, 37-38 in FIG. 2). The weighting unit preferably includes a second order polynomial weighting function (not to be confused with second order intermodulation). This weighting function has the following general formula:
W (f) _out = A · f ² + B · f + C
Here, W (f) _out is an output signal (in the frequency domain) of the weighting unit, f is a frequency, and A, B, and C are weighting parameters. The parameter adjustment unit 7 determines these weighting parameters using, for example, a well-known genetic optimization algorithm. Other optimization algorithms may be used, such as the well-known grid search algorithm.

  The adjustment of the weighting parameter of the weighting unit using the comparison test shown in FIG. 6 is not essential but optional. Instead, the weighting parameters can be determined in advance, thereby eliminating the optimization using the comparison test.

  The genetic optimization algorithm may include determining an initial value for the parameter and generating a number of parents each having a genetic structure corresponding to the initial parameter. Parents are ranked to produce the smallest difference signal DS according to the criteria for matching. The highest ranked parents are combined to form one or more children. The matched child is replaced with the parent ranked lower and becomes the new parent. This process is repeated by combining the parents with the highest rankings to further optimize the parameters. The various steps of the genetic algorithm are repeated until the optimal parameter that yields the local minimum of the difference signal is obtained.

  By using the comparative test shown in FIG. 6, the network model can be arbitrarily tuned, that is, the parameters of all the cable network models can be adjusted arbitrarily. In this case, the element model 3 is replaced with a network model (reference numeral 1 in FIG. 1 and reference numeral 1 ′ in FIG. 5), and the actual element 3 ′ is replaced with an actual network.

Specifically, the setting of this comparison test is an intermodulation signal (intermodulation component) which is a component, for example, PP, PQ,. . . , QQ, and relative contributions of PPP, PPQ, QQQ. According to a further aspect of the invention, the addition of these intermodulation components in the addition units 304, 305 shown in FIG. 5 is controlled by parameters. For second and third order intermodulation, two parameters A2 and A3 (and auxiliary variables k and n) are used.
IMS2 _TOTAL = (Σ [IMS2 ^{k / 2} ]) ^{2 / k} , k = (30−A2) / 10
IMS3 _TOTAL = (Σ [IMS3 ^{k / 2} ]) ^{2 / k} , k = (30−A3) / 10
Here, the default values are A2 = 10 (power addition) and A3 = 20 (amplitude addition), and as a result, k = 2 and n = 1. Σ represents the sum of all components (of the power spectrum) obtained by IMS2 and IMS3. IMS2 _TOTAL and IMS3 _TOTAL after the addition represent power spectra obtained by combining the second-order and third-order intermodulation components, respectively.

  However, it is desirable to adjust the addition parameters of the additional intermodulation components using a comparative test. In this case, the values of A2 and A3 are usually deviated from their initial values of 10 and 20, so as to better fit the network model. A well-known grid search algorithm can be used in the optimization process. This algorithm is more efficient than the genetic algorithm in the present case. However, other optimization algorithms may be used including genetic algorithms.

  An apparatus for determining signal quality of a cable network includes an input unit for inputting an appropriate input signal, a memory unit for storing a network model and its parameters, a processing unit for processing the input signal using the network model, and an input signal. A signal quality unit for determining the signal quality from the output signal and the intermodulation. This processing unit comprises a microprocessor and is connected to an input unit, a memory unit and a signal quality unit.

  Although the present invention has been described above with respect to a cable network such as CATV, the present invention is not limited to this, and can be applied to other electrical or optical networks such as a broadband network. Cable networks include, but are not limited to, coaxial cable networks, fiber optic networks, and fiber optic-coaxial cable (HFC) networks.

  The present invention is based on the consideration that the intermodulation and total noise levels in the network can be better estimated by separately determining the contribution of the various signal components to the intermodulation. The present invention further includes the consideration that frequency weighting of the amplified signal and intermodulation improves modeling accuracy, and the parameters of the element model that the genetic algorithm is used to determine the network model parameters, particularly the intermodulation contributions. And is considered to be used advantageously to optimize

  It should be noted that any terms in this specification should not be construed to limit the scope of the present invention. In particular, the terms “comprise (s)” and “comprising” are not meant to exclude elements not specifically mentioned. A single (circuit) element may be replaced with multiple (circuit) elements or their equivalents.

Those skilled in the art will appreciate that the present invention is not limited to the embodiments described above, and that many variations and additions can be made without departing from the scope of the invention as set forth in the appended claims. I will.

Claims (17)

A method for determining the signal quality of a network,
Providing a network model (1, 1 ′) comprising an interconnect (2) and at least one element model (3);
Providing an input signal (IS);
Determining an output signal (OS) using the input signal and the network model;
Determining a signal quality criterion (SNR, BER) using the input signal (IS) and the output signal (OS);
The network model is a model of a cable network such as a cable TV network,
The input signal (IS) includes component signal components (P, Q,...) Represented in a multi-frequency domain,
The output signal includes a representation of the output signal of the cable network in the frequency domain,
Determining the output signal (OS) comprises:
Simulating the behavior of the cable network in response to the input signal (IS) using the network model (1, 1 ');
Determining frequency domain intermodulation (IMS2, IMS3) by performing convolution of the spectrum of the constituent signal components (P, Q, ...);
Generating the signal quality criteria (SNR, BER) using the frequency domain intermodulation (IMS2, IMS3) and the output signal (OS). The method of claim 1, wherein
A method, characterized in that each of the spectral convolutions of the constituent signal components (P, Q, ...) is achieved by performing an inverse Fourier transform, a time-domain multiplication and a Fourier transform. The method according to claim 1 or 2, wherein
Method according to claim 1, characterized in that the network model (1, 1 ') is a frequency domain model. The method according to any one of claims 1 to 3,
The element model (3) includes a gain unit (31), a weighting unit (34, 35,...), And a weight unit (34), 35,. A method comprising at least one intermodulation unit (32). The method according to any one of claims 1 to 4,
Method according to claim 1, characterized in that the element model (3) comprises two intermodulation units (32, 33) for determining second-order and third-order intermodulation, respectively. The method according to claim 4 or 5, wherein
The at least one intermodulation unit (32, 33) has a first-order weighting unit (35, 36) that performs weighting of the input signal before determining the intermodulation in the preceding stage, and in the subsequent stage. And a second-order weighting unit (37, 38) for weighting the intermodulation. The method according to any one of claims 4 to 6,
Method according to claim 1, characterized in that said weighting unit (35, 36, 37, 38) comprises a quadratic weighting function. The method according to any one of claims 4 to 7,
Method according to claim 1, characterized in that the weighting unit (35, 36, 37, 38) comprises parameters determined by a genetic algorithm. A method according to any of claims 1 to 8,
The at least one intermodulation unit (32, 33) comprises intermodulation subunits (321, 322, ..., 331, 332, ...) that determine the intermodulation of the constituent signal components. And how to. A method according to any of claims 4 to 9,
The intermodulation units (32, 33) and / or the intermodulation subunits (321, 322,..., 331, 332,...) Of the constituent signal components (P, Q,. A method characterized in that it is configured to perform spectral convolution. A method according to any of claims 1 to 10,
The method wherein the signal quality criterion is a signal to noise ratio (SNR) and / or a bit error rate (BER). 12. A method according to any of claims 1 to 11,
A method characterized in that each signal component (P, Q,...) Represents the type of input signal. The method according to any of claims 1 to 12,
Measuring the signal generated by the actual element modeled in the element model (3);
Determining the difference between the calculated signal generated in the element model (3) and the measured signal;
Adjusting the parameters of the element model (3) based on the difference.   Computer program product for carrying out the method according to any of claims 1-12. An apparatus for determining signal quality of a network,
A memory unit storing a network model (1) comprising an interconnection (2) and at least one element model (3);
An input unit providing an input signal (IS);
A processing unit for determining an output signal (OS) using the input signal and the network model;
A signal quality unit for determining a signal quality standard using the input signal (IS) and the output signal (OS),
The network model stored in the memory unit is a model of a cable network such as a cable TV network,
The input unit is configured to receive an input signal (IS) including a component signal component (P, Q,...) Represented in a multi-frequency domain,
The processing unit is configured to determine an output signal (OS) including a representation of the output signal of the cable network in the frequency domain;
The processing unit is
Using the network model (1, 1 '), the behavior of the cable network according to the input signal (IS) is simulated,
Determine the frequency domain intermodulation (IMS2, IMS3) by performing a convolution of the spectrum of the constituent signal components (P, Q, ...);
The output signal (OS) is determined by generating the signal quality standard (SNR, BER) using the frequency domain intermodulation (IMS2, IMS3) and the output signal (OS). A device characterized by comprising. The apparatus of claim 15, wherein
The processing unit further achieves each of the spectral convolutions of the constituent signal components (P, Q,...) By performing an inverse Fourier transform, a time domain multiplication, and a Fourier transform. A device characterized in that it is configured. The apparatus according to claim 15 or 16,
An input that receives the signal from the actual element;
A parameter adjustment unit configured to adjust a parameter of the model (3) based on a difference between a calculation signal generated by the element model (3) and a measurement signal from the input; Equipment.

Images